nber working paper series labor market …

NBER WORKING PAPER SERIES

LABOR MARKET DISCRIMINATION ANDRACIAL DIFFERENCES IN PREMARKET FACTORS

Pedro CarneiroJames J. Heckman

Dimitriy V. Masterov

Working Paper 10068http://www.nber.org/papers/w10068

NATIONAL BUREAU OF ECONOMIC RESEARCH1050 Massachusetts Avenue

Cambridge, MA 02138October 2003

This research was supported by a grant from the American Bar Foundation and NIH R01-HD043411.Carneiro was supported by Fundacao Ciencia e Tecnologia and Fundacao Calouste Gulbenkian. We thankDerek Neal for comments and Maria Isabel Larenas, Maria Victoria Rodriguez and Xing Zhong for excellentresearch assistance. The views expressed herein are those of the authors and not necessarily those of theNational Bureau of Economic Research.

©2003 by Pedro Canreiro, James J. Heckman, and Dimitriy V. Masterov. All rights reserved. Short sectionsof text, not to exceed two paragraphs, may be quoted without explicit permission provided that full credit,including © notice, is given to the source.

Labor Market Discrimination and Racial Differences in Premarket FactorsPedro Carneiro, James J. Heckman, and Dimitriy V. MasterovNBER Working Paper No. 10068October 2003JEL No. J31

ABSTRACT

This paper examines minority-white wage gaps. Neal and Johnson (1996) show thatcontrolling for ability measured in the teenage years eliminates young adult wage gaps for all groupsexcept for black males, for whom they eliminate 70% of the gap. Their study has been faultedbecause minority children and their parents may have pessimistic expectations about receiving fairrewards for their skills and so they may invest less in skill formation. If this is the case,discrimination may still affect wages, albeit indirectly, though it would appear that any racialdifferences in wages are due to differences in acquired traits.

We find that gaps in ability across racial and ethnic groups open up at very early ages, longbefore child expectations are likely to become established. These gaps widen with age and schoolingfor Blacks, but not for Hispanics which indicates that poor schools and neighborhoods cannot be theprincipal factors affecting the slow black test score growth rate.

Test scores depend on schooling attained at the time of the test. Adjusting for racial andethnic differences in schooling attainment at the age the test is taken reduces the power of measuredability to shrink wage gaps for blacks, but not for other minorities.

The evidence from expectations data are mixed. Although all groups are quite optimisticabout future schooling outcomes, minority parents and children have more pessimistic expectationsabout child schooling relative to white children and their parents when their children are young. Atlater ages, expectations are more uniform across racial and ethnic groups. However, we also presentsome evidence that expectations data are unreliable and ambiguous.

We also document the presence of disparities in noncognitive traits across racial and ethnicgroups. These characteristics have been shown elsewhere to be important for explaining the labormarket outcomes of adults.

This evidence points to the importance of early (preschool) family factors and environmentsin explaining both cognitive and noncognitive ability differentials by ethnicity and race. Policies thatfoster both types of ability are far more likely to be effective in promoting racial and ethnic equalityfor most groups than are additional civil rights and affirmative action policies targeted at theworkplace.Pedro CarneiroDepartment of EconomicsUniversity College LondonGower StreetLondon WC1E 6BTUnited [email protected]

James J. HeckmanDepartment of EconomicsUniversity of Chicago1126 East 59th StreetChicago, IL 60637and [email protected]

Dimitriy V. MasterovIrving B. Harris SchoolCenter for Social Program EvaluationUniversity of Chicago1155 E. 60th Street, Room 038Chicago, IL [email protected]

1 Introduction

In spite of 40 years of civil rights and affirmative action policy, substantial gaps remain in the

market wages of African-American males and females compared to white males and females. There

are sizable wage gaps for Hispanics as well.1 Columns I of table 1 report the mean hourly log wage

gaps for a cohort of black and Hispanic males and females. These gaps are for a cohort of young

persons age 26-28 in 1990 from the National Longitudinal Survey of Youth of 1979, or NLSY79. We

follow the cohort for 10 years until it reaches age 36-38 in 2000. These gaps are not adjusted for

differences in schooling, ability, or other potential sources of racial and ethnic wage differentials.

Table 1 shows that, on average, black males earn 25% lower wages than white males in 1990.

Hispanic males earn 17.4% lower wages in the same year. Moreover, the gaps widen for males as

the cohort ages. The results for women reveal smaller gaps for blacks and virtually no gap at all

for Hispanic women.2 The gaps for women show no clear trend with age. Altonji and Blank (1999)

report similar patterns using data from the Current Population Survey (CPS).

These gaps are consistent with the claims of pervasive labor market discrimination against

many minorities, namely, that minority workers with the same ability and training as white workers

receive lower wages. However, there is another equally congruous explanation. Minorities may

bring less skill and ability to the market. Although there may be discrimination or disparity in

the development of these valuable skills, the skills may be rewarded equally across all demographic

groups in the labor market.

1The literature on African-American economic progress over the 20th century is surveyed in Heckman and Todd(2001).

2However, the magnitude (but not the direction) of the female gaps is less reliable, at least for black women. Neal(2003) shows that racial wage gaps for black women are underestimated by these types of regressions since they donot control for selective labor force participation. This same line of reasoning is likely to hold for Hispanic women.

1

These two interpretations of market wage gaps have profoundly different policy implications. If

persons of identical skill are treated differently on the basis of race or ethnicity, a more vigorous

enforcement of civil rights and affirmative action in the marketplace may be warranted. If the gaps

are due to unmeasured abilities and skills that people bring to the labor market, then a redirection

of policy towards fostering skills should be emphasized as opposed to a policy of ferreting out

discrimination in the workplace.

An important paper by Derek Neal and William Johnson (1996) sheds light on the relative

empirical importance of market discrimination and skill disparity in accounting for wage gaps by

race. Controlling for a measure of scholastic ability measured in the middle teenage years, they

substantially reduce but do not fully eliminate wage gaps for black males in 1990-1991. They

more than eliminate the gaps for black females. Columns II of tables 1 show our version of the

Neal-Johnson study,3 expanded to cover additional years. For black males, controlling for an early

measure of ability cuts the wage gap in 1990 by 76 percent. For Hispanic males, controlling for

ability essentially eliminates it. For women the results are even more striking. Wage gaps are

actually reversed, and controlling for ability leads to higher wages for minority females.

Following Neal and Johnson, these adjustments do not control for racial and economic differences

in schooling, occupational choice, or work experience. Neal and Johnson argue that racial and ethnic

differences in these factors may reflect responses to labor market discrimination and should not be

controlled for in estimating the full effect of race on wages since this may spuriously reduce estimated

wage gaps by introducing a proxy for discrimination into the control variables. They further argue

3We use a sample very similar to the one used in their study. It includes individuals born only in 1962-1964.This exclusion is designed to alleviate the effects of differential schooling at the test date on test performance andto ensure that the AFQT test is taken before the individuals enter the labor market (i.e., so that it is a premarketfactor).

2

that ability measured in the teenage years is a “premarket” factor, meaning that it is not affected by

expectations or actual experiences of discrimination in the labor market. We consider this claim and

argue that there is considerable arbitrariness in proclaiming what is or is not a “premarket” factor,

and such determinations matter greatly for the size of the adjusted wage gaps. When adjustments

for schooling attainment at the date of the test are made, the adjusted wage gaps rise.

Gaps in measured ability by ethnicity and race are indeed substantial. In figures 1A and 1B

we have plotted the ability distribution as measured by age-corrected AFQT4 for men and women,

respectively. These differences are not small. As noted by Herrnstein and Murray (1994), the ability

gaps are a major factor in accounting for a variety of racial and ethnic disparities in socioeconomic

outcomes. For example, Cameron and Heckman (2001) show that controlling for ability, blacks and

Hispanics are more likely to enter college than are whites at a time when wage premia for education

are rising considerably.5

The evidence in columns II of table 1 suggests that the major source of minority-majority

differences in wages is the disparity in characteristics that minorities bring to the market rather

than discrimination in the workplace. At first glance, the evidence in the table suggests that there

is no racial or ethnic disparity in market payments for comparable levels of skill for all but black

males.

Though the facts displayed in table 1 and figures 1A-B are provocative, they are controversial

4Age-corrected AFQT is the standardized residual from the regression of the AFQT score on age at the time of thetest dummy variables. AFQT is a subset of 4 out of 10 ASVAB tests used by the military for enlistment screeningand job assignment. It is the summed score from the word knowledge, paragraph comprehension, mathematicsknowledge, and arithmetic reasoning ASVAB tests.

5Urzua (2003) shows that this effect arises from greater minority enrollment in two-year colleges. Controlling forability, whites are more likely to attend and graduate from four year colleges. Using the Current Population Survey,Black and Sufi (2002) find that equating the family background of blacks and whites eliminates the black-white gapin schooling only at the bottom of the family background distribution. Furthermore, the gaps are eliminated in the1980s, but not in the 1990s.

3

for a number of reasons. The major points of contention are as follows.

1. The gaps in ability evident in figures 1A-B may stem from lowered academic effort in an-

ticipation of future discrimination in the labor market. If skills are not rewarded fairly, the

incentive to acquire them is diminished for those subject to prejudicial treatment. Discrim-

ination in the labor market might not only sap the incentives of children and young adults

to acquire skills and abilities, but it may also influence the efforts they exert in raising their

own offspring. This means that measured ability may not be a true premarket factor. Neal

and Johnson (1996) mention this qualification in their original paper and their critics have

subsequently reiterated it.

2. The gaps in ability may also be a consequence of adverse environments, and thus the appro-

priate policy for eliminating ability gaps is not apparent from table 1. Should policies focus

on early ages through enriched Head Start programs or on improving schooling quality and

reducing school dropout and repetition rates that plague minority children at later ages?

This paper provides answers to these questions. We show that:

(a) The evidence from data on parents’ and children’s expectations tells a mixed story. If the

rewards to schooling are lower for minorities, the return to schooling is lower and minority

expectations of schooling attainment should be lower. Minority child and parent expectations

measured when the children are 16 and 17 about the children’s schooling prospects are as

optimistic as white expectations, although actual schooling outcomes of whites and minorities

are dramatically different. Differential expectations at these ages cannot explain the gaps in

ability evident in figures 1A and 1B.

4

For children ages 14 and below, parent and child expectations about schooling are much

lower for blacks than for whites, though only slightly lower for Hispanics than for whites. All

groups are still rather optimistic in light of actual subsequent schooling attendance. At these

ages, differences in expectations across groups may lead to differential investments in skill

formation. Still, though lower expectations may be a consequence of perceived labor market

discrimination, they may also reflect child and parental perception of the lower endowments

possessed by minorities.

The fact that reported expectations are so inaccurate casts some doubt about the usefulness of

expectations elicited by questionnaires for testing the actual expectations governing behavior.

This is compounded by ambiguity of the source of the relatively pessimistic expectations.

(b) Ability gaps open up early, often by age 1 or 2, and as a result minorities enter school with

substantially lower measured ability than whites. The black-white ability gap widens as the

children get older and obtain more schooling, but the contribution of formal education to the

widening of the gap is small when compared to the size of the initial gap. There is a much

smaller widening of the Hispanic-white gap.

This evidence implies that school-based policies are unlikely to have substantial effects on

eliminating minority ability gaps. Factors that operate early in the life cycle of the child

are likely to have the greatest impact on ability. Although ability gaps between blacks and

whites widen with schooling, the effect of formal schooling on ability is much smaller for

blacks than for whites, and about the same for Hispanics and whites. The early emergence

of ability gaps indicates that child expectations can play only a limited role in accounting for

ability gaps since very young children are unlikely to have formed expectations about labor

5

market discrimination or to take decisions based on those expectations. However, parental

expectations of future discrimination may still play a role.

The early emergence of measured ability differentials also casts doubt on the empirical impor-

tance of the “stereotype threat” (see Steele and Aronson, 1998) as a major factor contributing

to black-white test score differentials. The literature on this topic finds that black college stu-

dents at selective colleges perform worse on tests when they are told that the outcomes will

be used in some way to measure black-white ability differences, i.e., that the test may be

used to confirm stereotypes about black-white ability differentials. However, the children in

our data are tested at a young age and are unlikely to be aware of stereotypes about minority

inferiority or be affected by the stereotype threat which has only been established for students

at elite college. In addition, large gaps in tests are also evident for Hispanics, a group for

whom the stereotype threat has not been documented.

(c) We find that differences in levels of schooling at the date the test is taken play a sizable role

in accounting for ability differentials. Adjusting for the schooling attainment of minorities at

the time that they take the test provides a potential qualification to the Neal and Johnson

finding. Part of the ability gap demonstrated by Neal and Johnson is due to differential

schooling attainment. An extra year of schooling has a greater impact on test scores for

whites and Hispanics than for blacks. Adjusting the test score for schooling disparity at the

date of the test raises the estimated wage gap and leaves more room for an interpretation of

labor market discrimination.

This finding does not necessarily overturn the conclusion of the Neal-Johnson analysis. At

issue is the source of the gap in schooling attainment at the date of the test. Our analysis

6

reveals that the amount of the wage gap that is explained by discrepancies in scholastic ability

depends on the age and grade completed at the date of measurement of ability. Tests adjusted

for schooling explain less of the black-white wage gap compared to the unadjusted tests. The

Neal-Johnson “pre-market” factors are a composite of ability and schooling, and are likely to

reflect both the life cycle experiences and the expectations of the child. To the extent that they

reflect expectations of discrimination as embodied in schooling that affects test scores, test

scores are contaminated by market discrimination and are not truly premarket factors. An

open question is how much of the gap in schooling is due to expectations about discrimination.

For black males premarket factors account for half of the black-white wage gap for males. We

argue that our adjustment is overly conservative because much of the gap in schooling and

ability opens up at an early age, before expectations of labor market discrimination can be a

plausible explanation. The adjustments for the effect of schooling on the test score have much

weaker effects for other demographic groups.

(d) A focus on cognitive skill gaps, while traditional (see, e.g., Jencks and Phillips, 1998), misses

important non-cognitive components of social and economic success. We show that non-

cognitive (i.e., behavioral) gaps also open up early. Previous work shows that they play

an important role in accounting for market wages. Policies that focus solely on improving

cognitive skills miss an important and promising determinant of socioeconomic success and

disparity that can be affected by policy (Carneiro and Heckman, 2003).

Section 2 presents evidence on the evolution of test score gaps over the lifecycle of the child.

Section 3 presents our evidence on how adjusting for schooling at the date of the test affects the

Neal-Johnson analysis, and how schooling affects test scores differentially for minorities. Section 4

7

discusses data on expectations. Section 5 presents evidence on noncognitive skills that parallels the

analysis of Section 2. Section 6 concludes.

2 Minority-White Differences in Early Test Scores and Early

Environments

The evidence presented in Section 1 suggests that a large fraction of the minority-white disparity

in labor market outcomes that is frequently attributed to discrimination may be due instead to

minority-white disparities in skill endowments. These skill endowments may be innate or acquired

during the lifetime of an individual. In this section we summarize evidence from the literature and

present original empirical work that demonstrates that minority-white cognitive skill gaps emerge

early and persist through childhood and the adolescent years.

Jencks and Phillips (1998) and Duncan and Brooks-Gunn (1997), among others, document

that the black-white test score gap is large for 3 and 4 year old children. Using the Children

of the NLSY79 (CNLSY) survey, a variety of studies show that even after controlling for many

variables like individual, family and neighborhood characteristics, the black-white test score gap

is still sizable. Using the Early Childhood Longitudinal Survey (ECLS), Fryer and Levitt (2002)

eliminate the black-white test score gap in math and reading for children at the time they are

entering kindergarten, although not in subsequent years. However, the raw test score gaps at

ages 3 and 4 are much smaller in ECLS than in CNLSY and other data sets that have been

used to study this issue. These studies also document that there are large black-white differences

in family environments. Ferguson (2002a) summarizes this literature and presents evidence that

8

black children in the ECLS come from much poorer and less educated families than white children,

and they are also more likely to grow up in single parent households. Studies summarized in

Ferguson (2002b) find that the achievement gap is high even for blacks and whites attending high

quality suburban schools.6 The common finding across these studies is that the black-white gap

in test scores is large and that it persists even after one controls for family background variables.

Children of different racial and ethnic groups grow up in strikingly different environments. Even

after accounting for these environmental factors in a correlational sense, test score gaps tend to

persist. Furthermore, they tend to widen with age and schooling: black children show lower ability

growth with schooling or age than white children.

In this paper we present some evidence from CNLSY.7 We have also examined the ECLS and

the Children of the Panel Study of Income Dynamics (CPSID) and found similar patterns. We

broaden previous analyses to include Hispanic-white differentials. Figures 2A-B shows the average

percentile PIAT Math8 scores for males and females in different age groups by race. Racial and

ethnic gaps are found as early as ages 5 and 6 (the earliest ages at which we can measure math

scores in CNLSY data).9 On average, black 5- and 6-year old boys are almost 18 percentile points

below white 5- and 6-year old boys (i.e., if the average white is at the 50th percentile of the test

score distribution, the average black is at the 32nd percentile of this distribution). The gap is a bit

smaller—16 percent—but still substantial for Hispanics. The finding is similar for black and Hispanic

women who exhibit gaps of about 14 percent relative to whites. These findings are duplicated for

6This is commonly referred to as the “Shaker Heights study,” although it analyzed many other similarneighborhoods.

7For a description of CNLSY and NLSY79 see BLS (2001).8The PIATMath is the abbreviation for Peabody Individual Achievement Test in Mathematics. This test measures

the child’s attainment in mathematics as taught in mainstream education. It consists of 84 multiple choice questionsof increasing difficulty, beginning with recognizing numerals and progressing to geometry and trigonometry.

9Instead of using raw scores or standardized scores we choose to use ranks, or percentiles, since test score scaleshave no intrinsic meaning. Our results are not sensitive to this procedure.

9

many other test scores and in other data sets, and are not altered if we use median test scores

instead of means. Furthermore, as shown in figures 3A-B, even when we use a test taken at earlier

ages, racial gaps in test scores can be found at ages 1 and 2, though not always for women.10 In

general, we find that the test score gaps emerge early and persist through adulthood.

For simplicity, we will focus on means and medians in this paper. However, figures 1A-B and

4A-B illustrate that there is considerable overlap in the distribution of test scores across groups in

recent generations. Many black and Hispanic children at ages 5 and 6 score higher on a math test

score than the average white child. Statements that we make about medians or means do not apply

to all persons in the distributions.

Figures 2A-B also shows that the black-white percentile PIAT Math score gap widens with age.

By ages 13 to 14, the average black is ranked more than 22 percentiles below the average white.

In fact, it is well documented that these gaps persist until adulthood and beyond. At 13 to 14

Hispanic boys are still almost 16 points below the average white. For black and Hispanic girls, the

gap widens to 21% and 16%, respectively.

In summary, when blacks and Hispanics enter the labor market, on average, they have a much

poorer set of skills than whites. Thus it is not surprising that their average labor market outcomes

are so much worse. Furthermore, these skill gaps emerge very early in the life-cycle, persist, and if

anything, widen for some groups. Initial conditions (i.e., early test scores) are very important as

skill begets skill.

Even after controlling for numerous environmental and family background factors, the racial

and ethnic test score gaps remain for many of these tests at ages 3 and 4, and for virtually all the

10Parts of the Body Test attempts to measure the young child’s receptive vocabulary knowledge of orally presentedwords as a means of estimating intellectual development. The interviewer names each of ten body parts and asksthe child to point to that part of the body.

10

tests at later ages. Figures 5A-B show that, even after adjusting for measures of family background,

such as family long-term or “permanent” income and mother’s education, the mother’s cognitive

ability (as measured by age-corrected AFQT), and a measure of home environment called the home

score,11 the black-white gap in percentile PIAT Math scores at ages 5-6 is almost 8%, and at ages

13-14 is close to 11% for boys. Hispanic-white differentials are reduced more by such adjustments,

falling to 7% at ages 5-6 and to 4% at ages 13-14 for boys. For other tests, differentials frequently

become positive or statistically insignificant. For girls, the gaps in PIAT Math are reduced to about

5% at ages 5-6 by the same adjustment, though the Hispanic gap falls to 4% while the black gap

rises to 6% by age 13-14. Appendix tables 1A-1B report that even after controlling for different

measures of home environments and child stimulation, the black-white test score gap persists even

though it drops considerably.12 Measured home and family environments play an important role in

the formation of these skills, although they are not the whole story.13 For females, both raw and

adjusted test score gaps are smaller than for males, but the overall story is the same.14

The evidence for Hispanics tells a different story.15 Early test scores for blacks and Hispanics

are similar, although Hispanics often perform slightly better. Figure 2A shows that for the PIAT

11The home score is the primary measure of the quality of a child’s home environment included in CNLSY. Itis composed of various measures of age-specific cognitive and emotional stimulation based on dichotomized andsummed responses from the mother’s self-report and interviewer observation. Some items included in the homescore are number of books, magazines, toys and musical recordings, how often mother reads to child, frequency offamily activities (eating, outings), methods of discipline and parenting, learning at home, television watching habitsand parental expectations for the child (chores, time use), home cleanliness and safety and types of mother-childinteractions.12Results for other tests and other samples can be found in http://lily.src.uchicago.edu/~dvmaster/Stanford.html.

Even though for some test scores early black-white test score gaps can be eliminated once we control for a largenumber of characteristics, it is harder to eliminate them at later ages. In the analysis presented here the mostimportant variable in reducing the test score gap is mother’s cognitive ability, as measured by the AFQT.13However the Home Score includes variables such as the number of books, which are clearly choice variables and

likely to cause problems in this regression. The variables with the largest effect on the minority-white test score gapare maternal AFQT and raw home score.14See http://lily.src.uchicago.edu/~dvmaster/Stanford.html.15We already saw that wage gaps are completely eliminated for Hispanics when we control for AFQT, while they

persist for blacks.

11

Math score the Hispanic-black gap is about 2 percentile points.16 This is much smaller than either

the black-white or the Hispanic-white gap. For the PIAT Math test, the black-white gap widens

dramatically, especially at later ages, but the Hispanic-white gap does not change substantially

with age. For other tests, even when there is some widening of the Hispanic-white gap with age,

it tends to be smaller than the widening in the black-white gap in test scores. In particular, when

we look at the AFQT scores displayed in figures 1A-B, and which are measured using individuals

at ages 16-23, Hispanics clearly have higher scores than blacks. In contrast, figures 4A-B show a

strong similarity between the math scores of blacks and Hispanics at ages 5 and 6, although there

are other tests where, even at these early ages, Hispanics perform significantly better than blacks.

When we control for the effects of home and family environments on test scores, the Hispanic-white

test score gap either decreases or is constant over time while the black-white test score tends to

widen with age.

Racial ability gaps open up very early. Home and family environments at early ages, and even

the mother’s behavior during pregnancy, are likely to play crucial roles in the child’s development,

and black children grow up in significantly more disadvantaged environments than white children.

Figure 6 shows the distributions of long-term or “permanent” family income for blacks, whites and

Hispanics. Minority children are much more likely to grow up in low income families than are white

children. Figure 7 shows the distribution of maternal education across race and ethnic groups.

Even though the overlap in the distributions is large, white children have more educated mothers

than do minority children. The distribution of maternal AFQT scores, shown in figure 8, is again

very different for minority and white children. Maternal AFQT is a major predictor of children’s

16The test score is measured in percentile rank The black-white gap is slightly below 18 while the Hispanic-whitegap is slightly below 16. This means that the black-Hispanic gap should be around 2.

12

test scores.17 Figure 9 document that white mothers are much more likely to read to their children

at young ages than are minority mothers, and we obtain similar results at young ages.18 Using

this reading variable and other variables in CNLSY such as number of books, magazines, toys

and musical recordings, family activities (eating, outings), methods of discipline and parenting,

learning at home, TV watching habits, parental expectations for the child (chores, time use), and

home cleanliness and safety, we can construct an index of cognitive and emotional stimulation—the

home score. Figure 10 shows that this index is always higher for whites than for minorities.19

Figure 11 shows that blacks are the most likely to grow up in broken homes. Hispanics are less

likely than blacks to grow up in a broken home, although they are much more likely to do so

than are whites. The research surveyed in Carneiro and Heckman (2003) suggests that enhanced

cognitive stimulation at early ages is likely to produce some achievement test gains in children from

disadvantaged environments, although long lasting effects of such interventions on IQ are not found.

The fact that racial and ethnic test score gaps open up early casts doubt on the empirical

importance of the stereotype threat. It is now fashionable in some circles to attribute gaps in black

test scores to racial consciousness on the part of black test takers stemming from the way test scores

are used in public discourse to describe minorities. Stereotype threats could not have been important

when blacks took the first IQ tests at the beginning of the twentieth century which documented

the racial differentials that gave rise to the stereotype. Yet racial IQ gaps are comparable across

time.20 Young children, like the ones studied in this paper, are unlikely to have the heightened racial

17For example, the correlation between percentile PIAT math score and age-corrected maternal AFQT is 0.4.18See the results for all ages in http://lily.src.uchicago.edu/~dvmaster/Stanford.html.19In http://lily.src.uchicago.edu/~dvmaster/Stanford.html, we document that both cognitive and emotional stim-

ulation indexes are always higher for whites than for blacks at all ages.20Murray (1999) reviews the evidence on the evolution of the black-white IQ gap. In the 1920s—a time when such

tests were much more unrealiable and black educational attainment much lower—the mean black-white differencewas 0.86 standard deviations. The largest black-white difference appears in the 1960s, with a mean black-whitedifference of 1.28 standard deviations. The difference ranges from a low of 0.82 standard deviations in the 1930s to

13

consciousness about tests and their social significance of the sort found by Steele and Aronson (1998)

in college students at a few elite universities. Moreover, sizable gaps are found for young Hispanic

males—a group for which the “stereotype” threat remains to be investigated.

3 The Differential Effect of Schooling on Test Scores

The previous section shows that cognitive test scores are correlated with home and family environ-

ments. We referenced research that suggests that test score gaps increase with age and schooling,

and presented fresh evidence on this issue. The research of Hansen, Heckman and Mullen (2003)

shows that the AFQT test scores used by Neal and Johnson are affected by schooling attainment

of individuals at the time they take the test. Therefore, one reason for the divergence of black and

white test scores over time may be differential schooling attainments. Figure 12 shows the schooling

completed at the test date for the six demographic groups used in the Neal and Johnson sample.

Blacks have slightly less completed schooling at test date than whites, but substantially more than

Hispanics. There are substantial schooling deficits for minorities.

Table 2 presents the effect of schooling at test date on AFQT scores for individuals in different

demographic groups in the NLSY, using a version of the nonparametric method developed in Hansen,

Heckman and Mullen (2003). Their method isolates the causal effect of schooling attained at the test

date on test scores controlling for unobserved factors that lead to selective differences in schooling

attainment. This table shows clearly that the effect of schooling on test scores is much higher for

whites and Hispanics than it is for blacks over most ranges of schooling. As a result, even though

Hispanics have fewer years of completed schooling at the time they take the AFQT test than blacks

1.12 standard deviations in the 1970s. However, none of the samples prior to 1960 are nationally representative, andthe samples were often chosen so as to effectively bias the black mean upward.

14

(see figure 12), on average Hispanics score better on the AFQT than do blacks.

There are different explanations for these facts. Carneiro and Heckman (2003) suggest that

one important feature of the learning process may be complementarity between endowment human

capital and learning.21 Higher levels of human capital raise the productivity of learning.22 If

minorities and whites start school with very different initial conditions (as documented in the

previous section), their learning paths can diverge dramatically over time. Another explanation

may be that blacks and non-blacks learn at different rates because blacks attend lower quality

schools than whites.

Currie and Thomas (2000) show that test score gains of participants in the Head Start program

tend to fade completely for blacks but not for whites. Their paper suggests that one reason may

be that blacks attend worse schools than whites, and therefore blacks are not able to maintain

initial test score gains. Both early advantages and disadvantages as well as school quality are

likely to be important factors in the human capital accumulation process. Therefore, differential

initial conditions and differential school quality may also be important determinants of the adult

black-white skill gap.

In light of the greater growth in test scores of Hispanics that are parallel to those of whites,

these explanations are not completely compelling. Hispanics start from similar initial disadvantages

in family environments and face school and neighborhood environments similar to those faced by

blacks.23 They also have early levels of test scores similar to those found in the black population.

What are the consequences of correcting for different levels of schooling at the test date? To

answer this question, we reanalyze Neal and Johnson’s 1996 study using AFQT scores corrected for

21For example, see the Ben-Porath (1967) model.22See the evidence in the paper by Heckman, Lochner and Taber (1998).23The evidence for CNLSY is presented on lily.src.uchicago.edu/~dvmaster/Stanford.html.

15

the race- or ethnicity-specific effect of schooling while equalizing the years of schooling attained at

the date of the test across all racial/ethnic groups. The results of this adjustment are given in table

3. This adjustment is equivalent to replacing each individual’s AFQT score by the score we would

measure if he or she would have stopped his or her formal education after eighth grade. (However,

the score is itself already affected by attendance in kindergarten, eight further years of schooling,

and any school quality differentials in those years.) In other words, we use “eighth grade” AFQT

scores for everyone. Since the effect of schooling on test scores is higher for whites than for blacks,

and whites have more schooling than blacks at the date of the test, this adjustment reduces the

test scores of whites much more. Although the black-white male wage gap is still cut in half when

we use this new measure of skill, the effect of AFQT on reducing the wage gap is weaker than in

the original Neal and Johnson (1996) study. The adjustment has little effect on the Hispanic-white

wage gap but a wage gap for black women emerges when using the schooling adusted measure.

This finding does not necessarily invalidate the Neal-Johnson study. It shows that schooling

can partially reduce the ability gap. It raises the larger question of what a “pre-market” factor

is. Neal and Johnson do not condition on schooling in explaining black-white wage gaps, arguing

that schooling is affected by expectations of adverse market opportunities facing minorities and

conditioning on such a contaminated variable would spuriously reduce the estimated wage gap. We

present direct evidence on this claim below.

Their line of argument is not entirely coherent. If expectations of discrimination affect schooling,

the very logic of their “pre-market” argument suggests that they should control for the impact of

schooling on test scores when using test scores to properly account for premarket factors. As we

have seen, when this is done, the wage gap (conditional on ability adjusted by schooling) widens

substantially for blacks. There is still little evidence of discrimination for Hispanic females, but the

16

evidence for Hispanic males comes close to demonstrating discrimination.

Implicitly, Neal and Johnson assume that schooling at the date of the test is not affected by

expectations of discrimination in the market while later schooling is. This distinction is arbitrary.

A deeper investigation of the expectation formation process and feedback is required. One practical

conclusion with important implications for the interpretation of the evidence is that the magnitude

of the wage gap one can eliminate by performing a Neal-Johnson analysis may depend on the age

at which the test is measured. The earlier the test is measured, the smaller the test score gap, and

the larger the fraction of the wage gap that is unexplained by the residual. Figures 13A-D show

how adjusting measured ability for schooling at the test at different levels affects the adjusted wage

gap for black males. For example, the log wage gap corresponding to grade of AFQT correction

equal to 11 is the log wage gap we obtain when use “eleventh grade” test scores. The later the

grade to which we adjust the test score, the lower the estimated gap because the greater the gap in

measured ability.

Finally we show that adjusting for expectations-contaminated completed schooling does not

operate in the fashion conjectured by Neal and Johnson. Table 4 shows that when we adjust wage

differences for completed schooling as well as schooling-adjusted AFQT, wage gaps widen. This runs

contrary to the simple intuition that schooling embodies expectations of market discrimination, so

that conditioning on it will eliminate wage gaps.24

Ours is a worst-case analysis for the Neal-Johnson study. If we assign all racial and ethnic

schooling differences to expectations of discrimination in the labor market, their results for blacks are

24The simple intuition, however, can easily be shown to be wrong so the evidence in these tables is not decisive onthe presence of discrimination in the labor market. The basic idea is that if both schooling and the test score arecorrelated with an unmeasured discrimination component in the error term, the bias for the race dummy may beeither positive or negative depending on the strength of the correlation among the contaminated variables and theircorrelation with the error term.

17

less sharp. Yet the evidence presented in Section 2 about the early emergence of ability differentials

is reinforced by the early emergence of differential grade repetition gaps by minorities (documented

by Cameron and Heckman, 2001). Most of the schooling gap at the date of the test emerges in the

early years, which child expectations are unlikely to be operative. Of course, one can always argue

that these early schooling and ability gaps are due to parental expectations of poor labor markets

for minority children. We examine data on parental expectations next.

4 The Role of Expectations

The argument that minorities perform worse on tests because they expect to be less well rewarded

in the labor market than whites for the same test score or schooling level is implausible because

expectations of labor market rewards are unlikely to affect the behavior of children as early as

ages 3 or 4 when test score gaps are substantial. The argument that minorities invest less in skills

because both minority children and minority parents have low expectations about their performance

in school and in the labor market has mixed empirical backing.

Data on expectations are hard to find, and when they are available they are difficult to interpret.

For example, in the NLSY97, black 17- and 18-year old report that the probability of dying next

year is 22% while for whites it is 16%. Both numbers are absurdly high. Minorities usually report

higher expectations than whites of committing a crime, being incarcerated and being dead next

year, and these adverse expectations may reduce their investment in human capital. Expectations

reported by parents and children for the child adolescent years for a variety of outcomes are given

in Table 5. Some estimates are plausible, while many others are not.

Schooling expectations measured in the late teenage years are very similar for minorities and

18

whites. They are slightly lower for Hispanics. Table 6A reports the mean expected probability

of being enrolled in school next year, for black, white and Hispanic 17- to 18-year old males.

Among those individuals enrolled in 1997, on average whites expect to be enrolled next year with

95.7% probability. Blacks expect that they will be enrolled next year with a 93.6% probability.

Hispanics expect to be enrolled with a 91.5% probability. If expectations about the labor market

were adverse for minorities, they should translate into adverse expectations for the child’s education.

Yet these data do not reveal this. Moreover, all groups substantially overestimate actual enrollment

probabilities. The difference in expectations between blacks and whites is very small, and is less

than half the difference in actual (realized) enrollment probabilities (81.9% for whites versus 76.4%

for blacks). The gap is wider for Hispanics. Table 6B reports parental schooling expectations for

white, black and Hispanic males for the same individuals used to present the numbers in Table 6A. It

shows that, conditional on being enrolled in 1997 (the year the expectation question is asked), black

parents expect their sons to be enrolled next year with a 90.9% probability, while for whites this

expectation is 95.4%. For Hispanics this number is lower (88.5%). Parents overestimate enrollment

probabilities for their sons, but black parents have lower expectations than white parents. For

females the racial and ethnic differences in parental expectations are smaller than those for males.25

For expectations measured at earlier ages the story is dramatically different. Figures 14A-B show

that, for the CNLSY, both black and Hispanic children and their parents have more pessimistic

expectations about schooling than white children, and more pessimistic expectations may lead to

lower investments in skills, less effort in schooling and lower ability. These patterns are also found

in the CPSID and ECLS.26 It is curious that for CNLSY teenagers expectations seem to converge

25See http://lily.src.uchicago.edu/~dvmaster/Stanford.html.26See http://lily.src.uchicago.edu/~dvmaster/Stanford.html.

19

at later ages (see figure 14C).

If the more pessimistic expectations of minorities are a result of market discrimination, then lower

investments in children that translate into lower levels of ability and skill at later ages are a result

of market discrimination. Ability would not be a premarket factor. However, lower expectations

for minorities may not be a result of discrimination but just a rational response to the fact that

minorities do not do as well in school as whites. This may be due to environmental factors unrelated

to expectations of discrimination in the labor market. Whether this phenomenon itself is a result

of discrimination is an open question. Expectation formation models are very complex and often

lead to multiple equilibria, and therefore they are difficult to test empirically.

Expectations are likely to play a very important role in the skill investment behavior of both

parents and children. Nevertheless, they are hard to measure, and data on them are often unreliable

or difficult to interpret. Across demographic groups, schooling expectations are in general unrealis-

tically optimistic although the degree of optimism varies by age. This is true at different ages, and

whether one measures expectations of parents or of their children. Since minorities perform much

more poorly than whites in terms of their schooling outcomes, relative expectations (expectation

minus actual probability of the event) tend to be much higher for minorities than for whites.

5 The Evidence on Non-Cognitive Skills

Controlling for scholastic ability in accounting for minority-majority wage gaps captures only part

of the endowment differences between groups but receives most of the emphasis in the literature on

black-white gaps (see Jencks and Phillips, 1998). An emerging body of evidence, summarized in

Bowles, Gintis and Osborne (2001) and Carneiro and Heckman (2003), documents that noncognitive

20

skills—motivation, self control, time preference, social skills—are important in explaining socioeco-

nomic success and differentials in socioeconomic success.

Some of the best evidence for the importance of noncognitive skills in the labor market is from

the GED (General Education Development) program. This program examines high school dropouts

to certify that they are equivalent to high school graduates. In its own terms, the GED program

is successful. Heckman and Rubinstein (2001) show that GED recipients and ordinary high school

graduates who do not go on to college have the same distribution of AFQT scores (the same test

that is graphed in figure 1). They are psychometrically equated. Yet GED recipients earn the wages

of high school dropouts with the same number of years of completed schooling. They are more likely

to quit their jobs, engage in fighting or petty crime, or to be discharged from the military, than

are high school graduates who do not go on to college or other high school dropouts (see Heckman

and Rubinstein, 2001). Intelligence alone is not sufficient for socioeconomic success. Minority-white

gaps in noncognitive skills open up early and widen over the lifecycle.

The CNLSY has lifecycle measures of noncognitive skills. Mothers are asked age specific ques-

tions about the anti-social behavior of their children such as aggressiveness or violent behavior,

cheating or lying, disobedience, peer conflicts and social withdrawal. The answers to these ques-

tions are grouped in different indices.27 Figures 15A-B shows that there are important racial and

ethnic gaps in anti-social behavior index that emerge in early childhood. By ages 5 and 6, the

average black is roughly 10 percentile points above the average white in the distribution of this

27The children’s mothers were asked 28 age-specific questions about frequency, range and type of specific behaviorproblems that children age four and over may have exhibited in the previous three months. Factor analysis was usedto determine six clusters of questions. The responses for each cluster were then dichotomized and summed. TheAntisocial Behavior index we use in this paper consists of measures of cheating and telling lies, bullying and crueltyto others, not feeling sorry for misbehaving, breaking things deliberately (if age is less than 12), disobedience atschool (if age is grater than 5), and trouble getting along with teachers (if age is greater than 5).

21

score (the higher the score, the worse the behavior).28 The results shown in Figures 16A-B— where

we adjust the gaps by permanent family income, mother’s education and age-corrected AFQT and

home score—also show large reductions.29

In Section 2 we documented that minority and white children face sharp differences in family

and home environments while growing up. The evidence presented in this section shows that these

early environmental differences can account (in a correlational sense) for most of the minority-white

gap in noncognitive skills, as measured in CNLSY.

Carneiro and Heckman (2003) document that noncognitive skills are more malleable than cogni-

tive skills. In their paper they report that noncognitive skill gaps can be eliminated by equalization

of family and home environments, while cognitive skill gaps cannot, even though they can be sharply

reduced. The largest effects of interventions in childhood and adolescence are on noncognitive skills.

Improvements in these skills produce better labor market outcomes, less engagement in criminal

activities and other risky behavior. This is an avenue for policy that warrants much greater atten-

tion.

6 Summary and Conclusion

This paper discusses the sources of wage gaps between minorities and whites. For all minorities but

black males, adjusting for the ability that minorities bring to the market eliminates wage gaps. The

major source of economic disparity by race and ethnicity in U.S. labor markets is in endowments,

not in payments to endowments.

28In http://lily.src.uchicago.edu, we show that these differences are statistically strong. Once we control for familyand home environments, gaps in most behavioral indices disappear.29See Appendix Tables 2A-2B for the effect of adjusting for other environmental characteristics on the anti-social

behavior score.

22

This evidence suggests that strengthened civil rights and affirmative action policies targeted at

the labor market are unlikely to have much effect on racial and ethnic wage gaps, except possibly for

those specifically targeted toward black males. Policies that foster endowments have much greater

promise. On the other hand, this paper does not provide any empirical evidence on whether or not

the existing edifice of civil rights and affirmative action legislation should be abolished. All of our

evidence on wages is for an environment where affirmative action laws and regulations are in place.

Minority deficits in cognitive and noncognitive skills emerge early and widen. Unequal school-

ing, neighborhoods and peers may account for this differential growth in skills, but the main story

in the data is not about growth rates but rather about the size of early deficits. Hispanic children

start with cognitive and noncognitive deficits similar to those of black children. They also grow up

in similar disadvantaged environments, and are likely to attend schools of similar quality. Hispanics

have substantially less schooling than blacks. Nevertheless, the ability growth by years of schooling

is much higher for Hispanics than blacks. By the time they reach adulthood, Hispanics have signif-

icantly higher test scores than blacks. Conditional on test scores, there is no Hispanic-white wage

gap. Our analysis of the Hispanic data illuminates the traditional study of black-white differences

and casts doubt on many conventional explanations of these differences since they do not apply to

Hispanics who also suffer from many of the same disadvantages. The failure of the Hispanic-white

gap to widen with schooling or age casts doubt on poor schools and bad neighborhoods as the

reasons for the slow growth rate in black test scores, since Hispanics experience the same kinds of

poor schools and bad neighborhoods experienced by blacks. Deficits in noncognitive skills can be

explained (in a statistical sense) by adverse early environments; deficits in cognitive skills are less

easily eliminated by the same factors.

Neal and Johnson (1996) were the first to document that endowments acquired before people

23

enter the market explain most of the minority-majority wage gap. They use an ability test taken

in the teenage years as a measure of endowment unaffected by discrimination. In this paper,

we note that the distinction between “premarket” factors unaffected by expectations of market

discrimination and “market” factors so affected is an arbitrary one. Occupational choice is likely

affected by discrimination. The proper treatment of schooling is less clear-cut. Neal and Johnson

purposely omit schooling in adjusting for racial and ethnic wage gaps, arguing that schooling choices

are potentially contaminated by expectations of labor market discrimination. Yet they do not

adjust their measure of ability by the schooling attained at the date of the test, which would be the

appropriate correction if their argument were correct. The literature establishes that their measure

of ability is affected by this measure of schooling.

Adjusting wage gaps by completed schooling and the adjusted test widens wage gaps for all

groups, but has especially strong effects for blacks. At issue is how much of the difference in

schooling at the date of the test is due to expectations of labor market discrimination and how much

is due to adverse early environments. While this paper does not settle this question definitively, test

score gaps emerge early and are more plausibly linked to adverse early environments. The lion’s

share of the ability gaps at the date of the test emerge very early, before children can have clear

expectations about their labor market prospects.

The emergence of test score gaps in young children casts serious doubt on the importance of

“stereotype threats” in accounting for poorer black test scores. It is implausible that young minority

test takers have the social consciousness assumed in the stereotype literature.

Moreover, gaps in test scores of the magnitude found in recent studies were found in the earliest

tests before the results of testing were disseminated. Effective social policy designed to eliminate

racial and ethnic inequality for most minorities should focus on eliminating skill gaps, not on

24

discrimination in the workplace of the early Twenty-First Century. Interventions targeted at adults

are much less effective and do not compensate for early deficits. Early interventions aimed at young

children hold much greater promise than strengthened legal activism in the workplace.30

30See Carneiro and Heckman (2003) for the evidence on early interventions and later remedial interventions.

25

References

[1] Altonji, Joseph and Rebecca Blank, (1999). “Gender and Race in the Labor Market,” in O.

Ashenfelter and D. Card, Handbook of Labor Economics, Volume 3C. (New York, NY: Elsevier

Science Press).

[2] Ben-Porath, Yoram (1967). “The Production of Human Capital and the Life Cycle Earnings,”

Journal of Political Economy, 75(4, part 1): 352-365.

[3] Black, Sandra E. and Amir Sufi (2002). “Who Goes to College? Differential Enrollment by

Race and Family Background,” NBER Working Paper no. w9310.

[4] Bureau of Labor Statistics (BLS) (2001). NLS Handbook 2001. Washington, D.C.: U.S. De-

partment of Labor.

[5] Cameron, Steve and James J. Heckman, (2001).“The Dynamics of Educational Attainment for

Blacks, Whites and Hispanics,” Journal of Political Economy 109(3) (2001), 455-499.

[6] Carneiro, Pedro and James J. Heckman, (2003). “Human Capital Policy,” Forthcoming in J.

Heckman and A. Krueger (eds.), Inequality in America: What Role for Human Capital Policy?

(Cambridge, MA: MIT Press).

[7] Cawley, John, James J. Heckman and Edward Vytlacil, (1999). “On Policies to Reward the

Value Added by Educators,” Review of Economics and Statistics, 81(4): 720-728.

[8] Currie, Janet, and Duncan Thomas. (2000). “School Quality and the Longer-Term Effects of

Head Start,” Journal of Human Resources, 35(4): 755-74.

26

[9] Duncan, Greg, and Jeanne Brooks-Gunn. (1997). Consequences of Growing Up Poor. (New

York: Russell Sage).

[10] Ferguson, Ronald (2002a), “Why America’s Black-White School Achievement Gap Persists,”

Unpublished Manuscript, Harvard University, 2002.

[11] Ferguson, Ronald (2002b), “What Doesn’t Meet the Eye: Understanding and Addressing

Racial Disparities in High Achieving Suburban Schools,” Special Edition, Policy Issues Re-

port, (Naperville, IL: North Central Regional Educational Laboratory).

[12] Fryer, Roland, and Steven Levitt. (2002). “Understanding the Black-White Test Score Gap in

the First Two Years of School,” NBER working paper no. 8975.

[13] Hansen, Karsten, James J. Heckman and Kathleen Mullen, (2003). “The Effect of Schooling

and Ability on Achievement Test Scores,” Forthcoming, Journal of Econometrics.

[14] Heckman, James J., Lance Lochner and Christopher Taber, (1998). “Explaining Rising Wage

Inequality: Explorations with a Dynamic General Equilibrium Model of Labor Earnings with

Heterogeneous Agents,” Review of Economic Dynamics, 1(1): 1-58.

[15] Heckman, James J. and Yona Rubinstein, (2001). “The Importance of Noncognitive Skills:

Lessons from the GED Testing Program,” American Economic Review, 91(2): 145-49.

[16] Heckman, J., J. Hsee and Y. Rubinstein (2003). “The GED is a ‘Mixed Signal’: The Effect of

Cognitive and Noncognitive Skills on Human Capital and Labor Market Outcomes,” working

paper, University of Chicago.

27

[17] Heckman, James J. and Petra Todd, (2001). “Understanding the Contribution of Legislation,

Social Activism, Markets and Choice to the Economic Progress of African Americans in the

Twentieth Century,” Unpublished manuscript, American Bar Foundation, Chicago, IL.

[18] Herrnstein, Richard and Charles Murray, (1994). The Bell Curve : Intelligence and class struc-

ture in American life. (New York: Free Press).

[19] Jencks, Christopher and Meredith Phillips (1998),The Black-White Test Score Gap. (Washing-

ton, D.C.: The Brookings Institution).

[20] Murray, Charles (1999). “The Secular Increase in IQ and Longitudinal Changes in the Magni-

tude of the Black-White Difference: Evidence from the NLSY,” paper presented to the Behavior

Genetics Association Meeting.

[21] Neal, Derek (2003), “The Measured Black-White Wage Gap among Women is Too Small,”

working paper, University of Chicago.

[22] Neal, Derek and William Johnson (1996), “The Role of Premarket Factors in Black-White

Wage Differences,” Journal of Political Economy 104(5): 869-95.

[23] Steele, Claude and Joshua Aronson (1998), “Stereotype Threat and the Test Performance of

Academically Successful African Americans,” in C. Jencks and M. Phillips, eds., The Black-

White Test Score Gap, (Washington, DC: The Brookings Institution), pp. 401-427..

[24] Urzua, Sergio (2003), “The Educational White-Black Gap: Evidence on Years of Schooling”,

Unpublished manuscript, University of Chicago, Department of Economics.

28

YearI II I II I II I II I II I II I II I II

Black -0.250 -0.060 -0.251 -0.082 -0.302 -0.113 -0.282 -0.104 -0.286 -0.093 -0.373 -0.149 -0.333 -0.069 -0.325 -0.089(0.028) (0.030) (0.028) (0.030) (0.029) (0.030) (0.028) (0.030) (0.031) (0.033) (0.032) (0.034) (0.034) (0.035) (0.035) (0.035)

Hispanic -0.174 -0.035 -0.113 0.020 -0.146 -0.014 -0.159 -0.027 -0.143 0.005 -0.186 -0.031 -0.195 -0.006 -0.215 -0.053(0.032) (0.032) (0.032) (0.033) (0.033) (0.032) (0.032) (0.032) (0.036) (0.036) (0.038) (0.038) (0.040) (0.038) (0.040) (0.038)

Age - 0.050 - 0.030 - 0.038 - 0.030 - 0.023 - 0.020 - 0.014 - 0.008- (0.014) - (0.014) - (0.014) - (0.014) - (0.016) - (0.016) - (0.017) - (0.017)

AFQT - 0.183 - 0.161 - 0.179 - 0.172 - 0.188 - 0.216 - 0.254 - 0.241- (0.013) - (0.013) - (0.013) - (0.013) - (0.014) - (0.015) - (0.015) - (0.015)

AFQT2 - -0.022 - -0.007 - -0.001 - 0.002 - 0.014 - 0.021 - 0.032 - 0.023- (0.011) - (0.011) - (0.011) - (0.011) - (0.012) - (0.013) - (0.013) - (0.013)

Intercept 2.375 0.957 2.372 1.463 2.404 1.202 2.423 1.431 2.458 1.652 2.533 1.756 2.589 1.960 2.629 2.224(0.017) (0.385) (0.017) (0.406) (0.017) (0.422) (0.017) (0.432) (0.018) (0.488) (0.019) (0.540) (0.020) (0.594) (0.020) (0.621)

N 1538 1505 1553 1514 1536 1503 1542 1504 1522 1485 1554 1519 1494 1462 1438 1404

Year:I II I II I II I II I II I II I II I II

Black -0.172 0.041 -0.200 0.030 -0.201 0.010 -0.167 0.093 -0.148 0.099 -0.147 0.132 -0.201 0.071 -0.200 0.069(0.031) (0.033) (0.032) (0.035) (0.031) (0.033) (0.035) (0.037) (0.035) (0.037) (0.035) (0.036) (0.034) (0.036) (0.036) (0.038)

Hispanic -0.003 0.154 -0.017 0.153 -0.059 0.114 0.009 0.198 -0.018 0.170 -0.006 0.193 -0.069 0.151 -0.064 0.149(0.035) (0.035) (0.037) (0.038) (0.036) (0.035) (0.039) (0.039) (0.040) (0.040) (0.041) (0.039) (0.039) (0.039) (0.041) (0.041)

Age - 0.010 - 0.038 - 0.016 - 0.016 - 0.008 - -0.009 - 0.013 - -0.018- (0.015) - (0.017) - (0.016) - (0.017) - (0.018) - (0.018) - (0.017) - (0.018)

AFQT - 0.217 - 0.234 - 0.229 - 0.271 - 0.267 - 0.283 - 0.274 - 0.273- (0.016) - (0.018) - (0.017) - (0.018) - (0.019) - (0.018) - (0.018) - (0.018)

AFQT2 - 0.005 - 0.000 - -0.001 - -0.012 - -0.024 - 0.005 - 0.000 - -0.008- (0.014) - (0.014) - (0.014) - (0.015) - (0.016) - (0.015) - (0.015) - (0.015)

Intercept 2.141 1.750 2.175 0.982 2.193 1.615 2.174 1.558 2.218 1.858 2.246 2.412 2.311 1.724 2.339 2.867(0.019) (0.420) (0.020) (0.467) (0.019) (0.458) (0.021) (0.520) (0.022) (0.555) (0.022) (0.582) (0.021) (0.603) (0.022) (0.663)

N 1356 1325 1335 1299 1317 1278 1319 1281 1318 1288 1381 1344 1370 1329 1316 1276

Age-corrected AFQT is the standardized residual from the regression of the AFQT score on age at the time of the test dummy variables. AFQT is a subset of 4 out of 10 ASVAB tests used by the military for enlistment screening and job assignment. It is the summed score from the word knowledge, paragraph comprehension, mathematics knowledge, and arithmetic reasoning ASVAB tests. All wages are in 1993 dollars. The coefficients on the AFQT variables represent the effect of a one standard deviation increase in the score on the log hourly wage. Since the wage is measured in log points, the gaps for blacks and hispanics correspond approximately to percentage point differences relative to the white mean; that is, the black-white gap of -.25 in 1990 corresponds to 25% lower wages for blacks in that year.

1991 1992

1996 1998

1990

2000

Panel B. NLSY Women Born After 19611996 20001993 1994 1998

Table 1Change in the Black-White Log Wage Gap Induced by Controlling for Age-Corrected AFQT in 1990-2000

Panel A. NLSY Men Born After 19611990 1991 1992 1993 1994

0.1

.2.3

Den

sity

−4 −3 −2 −1 0 1 2 3 4Standardized Score

White Black Hispanic

Age−corrected AFQT is the standardized residual from the regression of the AFQT score on age at thetime of the test dummy variables. AFQT is a subset of 4 out of 10 ASVAB tests used by the militaryfor enlistment screening and job assignment. It is the summed score from the word knowledge, paragraphcomprehension, mathematics knowledge, and arithmetic reasoning ASVAB tests.

NLSY79 Males Born After 1961

Figure 1ADensity of Age−Corrected AFQT

0.1

.2.3

Den

sity

−4 −3 −2 −1 0 1 2 3 4Standardized Score



NLSY79 Females Born After 1961

Figure 1BDensity of Age−Corrected AFQT

40

45

50

55

60

Ave

rage

Per

cent

ile S

core

5−6 7−8 9−10 11−12 13−14Child’s Age

Hispanic Black White

This test measures the child’s attainment in mathematics as taught in mainstream education. It consists of 84multiple−choice questions of increasing difficulty, beginning with recognizing numerals and progressing togeometry and trigonometry. The percentile score was calculated separately for each sex at each age.span

Children of NLSY79 Males

Figure 2APercentile PIAT Math Score By Race and Age Group

40

45

50

55

60

Ave

rage

Per

cent

ile S

core

5−6 7−8 9−10 11−12 13−14Child’s Age


This test measures the child’s attainment in mathematics as taught in mainstream education. It consists of 84multiple−choice questions of increasing difficulty, beginning with recognizing numerals and progressing togeometry and trigonometry. The percentile score was calculated separately for each sex at each age.span

Children of NLSY79 Females

Figure 2BPercentile PIAT Math Score By Race and Age Group

30

35

40

45

50

Ave

rage

Per

cent

ile S

core

1 2 3Child�s Age


This test attempts to measure the young child�s receptive vocabulary knowledge of orally presented words as a meansof estimating intellectual development. The interviewer names each of ten body parts and asks the child to point to thatpart of the body. The score is computed by summing the number of correct responses. The percentile score was calculatedseparately for each sex at each age.


Figure 3AAverage Percentile Parts of the Body Score By Race and Age

20

30

40

50

60

Ave

rage

Per

cent

ile S

core

1 2 3Child�s Age


This test attempts to measure the young child�s receptive vocabulary knowledge of orally presented words as a meansof estimating intellectual development. The interviewer names each of ten body parts and asks the child to point to thatpart of the body. The score is computed by summing the number of correct responses. The percentile score was calculatedseparately for each sex at each age.


Figure 3BAverage Percentile Parts of the Body Score By Race and Age

0.0

05.0

1.0

15D

ensit

y

0 10 20 30 40 50 60 70 80 90 100Percentile Score


This test measures the child’s attainment in mathematics as taught in mainstream education. It consists of 84multiple−choice questions of increasing difficulty, beginning with recognizing numerals and progressing togeometry and trigonometry. The percentile score was calculated separately for each sex at each age.

CNLSY 79 Males

Figure 4ADensity of Percentile PIAT Math Scores at Ages 5−6

.002

.004

.006

.008

.01

.012

Den

sity

0 10 20 30 40 50 60 70 80 90 100Percentile Score


This test measures the child’s attainment in mathematics as taught in mainstream education. It consists of 84multiple−choice questions of increasing difficulty, beginning with recognizing numerals and progressing togeometry and trigonometry. The percentile score was calculated separately for each sex at each age.

CNLSY 79 Females

Figure 4BDensity of Percentile PIAT Math Scores at Ages 5−6

40

45

50

55

Ave

rage

Per

cent

ile S

core

5−6 7−8 9−10 11−12 13−14Child’s Age


Adjusted by permanent family income, mother’s education and age−corrected AFQT, and home score.Adjusted indicates that we equalized the family background characteristics across all race groups by setting them atthe mean to purge the effect of family environment disparities. Permanent income is constructed by taking the averageof annual family income discounted to child’s age 0 using a 10% discount rate. Age−corrected AFQT is the standardizedresidual from the regression of the AFQT score on age at the time of the test dummy variables. Home score is anindex of quality of the child’s home environment.


Figure 5AAdjusted Percentile PIAT Math Score By Race and Age Group

44

46

48

50

52

54

Ave

rage

Per

cent

ile S

core

5−6 7−8 9−10 11−12 13−14Child’s Age


Adjusted by permanent family income, mother’s education and age−corrected AFQT, and home score.Adjusted indicates that we equalized the family background characteristics across all race groups by setting them atthe mean to purge the effect of family environment disparities. Permanent income is constructed by taking the averageof annual family income discounted to child’s age 0 using a 10% discount rate. Age−corrected AFQT is the standardizedresidual from the regression of the AFQT score on age at the time of the test dummy variables. Home score is anindex of quality of the child’s home environment.


Figure 5BAdjusted Percentile PIAT Math Score By Race and Age Group

0.1

.2.3

.4.5

Den

sity

0 5 10 15Log of Permanent Income


Permanent income is constructed by taking the average of annual familyincome discounted to child’s age 0 using a 10% discount rate.

CNLSY 79 Males and Females

Figure 6Density of Log Permanent Income

0.0

5.1

.15

Den

sity

0 5 10 15 20Highest Grade Completed


CNLSY79 Males and Females

Figure 7Density of Mother’s Highest Grade Completed

0.1

.2.3

.4.5

Den

sity

−4 −2 0 2 4Standardized Score




Figure 8Density of Mother’s Age−Corrected AFQT

0 .2 .4 .6 0 .2 .4 .6

White

Black

Hispanic

White

Black

Hispanic

Male Female

Never Several Times a Year Several Times a Month Once A Week About 3 Times a Week Every Day

Fraction

The height of the bar is produced by dividing the number of people who report falling in a particular reading frequency cellby the total number of people in their race−sex group

CNLSY79

Figure 9How Often Mother Reads to Child at Age 2

by Race and Sex

0.0

02.0

04.0

06.0

08.0

1D

ensit

y

0 100 200 300Raw Home Score


Home Score is an index of the quality of a child’s home environment included in CNLSY79.It consists of various measures of age−specific cognitive and emotional stimulation basedon dichotomized and summed responses to mother’s self−report and interviewer observation.


Figure 10Density of Home Score

0

.1

.2

.3

.4

.5

.6

Frac

tion

With

Bro

ken

Hom

e

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14Child’s Age

Hispanic Children Black Children White Children

Note: We define a child living in a broken home as a child not living in a household with two biological parents,or living with the mother and her common law partner


Figure 11Fraction of Children Who Live in a Broken Home by Age and Race

0 .1 .2 .3 .4 0 .1 .2 .3 .4

Hispanic

Black

White

Hispanic

Black

White

Male Female

Less Than or Equal To 9 Years 10 Years 11 Years 12 Years 13−14 Years

Fraction

The height of the bar is produced by dividing the number of people who report falling in a particular education cellby the total number of people in their race−sex group

NLSY79 Born After 1961

Figure 12 Highest Category of Schooling Completed at Test Date

by Race, Sex and Age

Schooling At Test Date Male Female Male Female Male Female9 12.68 14.66 5.21 5.95 9.80 11.22Std. Err. (1.51) (1.70) (1.58) (1.74) (1.90) (1.79)N 378 343 234 234 167 17910 16.94 16.12 9.22 5.65 16.33 14.99Std. Err. (1.52) (1.68) (1.53) (1.73) (1.94) (1.85)N 377 368 282 237 161 16911 22.02 18.55 8.87 10.62 18.67 16.83Std. Err. (1.54) (1.72) (1.58) (1.75) (2.11) (2.01)N 366 327 247 243 120 12212 23.12 21.05 11.96 11.12 21.29 16.46Std. Err. (1.49) (1.64) (1.58) (1.70) (2.08) (1.95)N 630 758 322 393 171 19813-15 26.60 24.37 15.37 14.23 23.96 18.92Std. Err. (1.73) (1.83) (2.23) (2.00) (2.82) (2.55)N 266 326 98 180 72 8116 or More 29.02 25.71 28.43 22.66 33.11 31.18Std. Err. (2.13) (2.24) (3.56) (2.99) (4.70) (4.03)N 108 103 27 34 17 22

Mean 52.50 51.95 36.79 37.44 38.45 36.87Std. Dev (19.11) (17.81) (17.76) (16.15) (19.17) (17.80)N 165 110 172 98 154 137

Table 2

Panel A. Effect of Schooling on Raw AFQTaNLSY79

AFQT and Schooling


a Baseline is less than or equal to 8 years of schooling. N corresponds to the number ofobservations in each race-gender-schooling at test date cell.

Panel B. Mean Schooling-Corrected AFQT For Those With 8 Years of Schooling


Black -0.250 -0.133 -0.251 -0.149 -0.302 -0.180 -0.282 -0.171 -0.286 -0.165 -0.373 -0.230 -0.333 -0.160 -0.325 -0.172(0.028) (0.029) (0.028) (0.029) (0.029) (0.029) (0.028) (0.029) (0.031) (0.031) (0.032) (0.032) (0.034) (0.034) (0.035) (0.033)

Hispanic -0.174 -0.070 -0.113 -0.013 -0.146 -0.044 -0.159 -0.058 -0.143 -0.029 -0.186 -0.067 -0.195 -0.047 -0.215 -0.088(0.032) (0.032) (0.032) (0.032) (0.033) (0.032) (0.032) (0.032) (0.036) (0.036) (0.038) (0.038) (0.040) (0.038) (0.040) (0.038)

Age - 0.065 - 0.043 - 0.055 - 0.045 - 0.040 - 0.039 - 0.036 - 0.029- (0.014) - (0.015) - (0.015) - (0.015) - (0.016) - (0.017) - (0.017) - (0.017)

AFQT - 0.153 - 0.131 - 0.155 - 0.144 - 0.159 - 0.184 - 0.221 - 0.211- (0.013) - (0.013) - (0.013) - (0.013) - (0.014) - (0.015) - (0.015) - (0.015)

AFQT2 - 0.001 - 0.009 - 0.015 - 0.017 - 0.028 - 0.031 - 0.040 - 0.036- (0.010) - (0.010) - (0.010) - (0.010) - (0.011) - (0.012) - (0.012) - (0.012)

Constant 2.375 0.540 2.372 1.085 2.404 0.732 2.423 0.982 2.458 1.119 2.533 1.140 2.589 1.172 2.629 1.425(0.017) (0.392) (0.017) (0.412) (0.017) (0.426) (0.017) (0.437) (0.018) (0.493) (0.019) (0.547) (0.020) (0.603) (0.020) (0.628)

N 1538 1505 1553 1514 1536 1503 1542 1504 1522 1485 1554 1519 1494 1462 1438 1404


Black -0.172 -0.045 -0.200 -0.066 -0.201 -0.083 -0.167 -0.020 -0.148 -0.014 -0.147 0.025 -0.201 -0.043 -0.200 -0.041(0.031) (0.031) (0.032) (0.033) (0.031) (0.031) (0.035) (0.035) (0.035) (0.035) (0.035) (0.034) (0.034) (0.034) (0.036) (0.035)

Hispanic -0.003 0.116 -0.017 0.107 -0.059 0.069 0.009 0.145 -0.018 0.119 -0.006 0.149 -0.069 0.098 -0.064 0.096(0.035) (0.035) (0.037) (0.037) (0.036) (0.035) (0.039) (0.039) (0.040) (0.040) (0.041) (0.039) (0.039) (0.038) (0.041) (0.040)

Age - 0.015 - 0.042 - 0.021 - 0.024 - 0.015 - -0.003 - 0.019 - -0.010- (0.016) - (0.017) - (0.016) - (0.018) - (0.018) - (0.018) - (0.017) - (0.018)

AFQT - 0.188 - 0.197 - 0.187 - 0.221 - 0.221 - 0.245 - 0.228 - 0.235- (0.016) - (0.017) - (0.016) - (0.018) - (0.018) - (0.018) - (0.017) - (0.018)

AFQT2 - 0.010 - 0.009 - 0.010 - 0.006 - -0.008 - 0.022 - 0.017 - 0.005- (0.013) - (0.014) - (0.013) - (0.015) - (0.015) - (0.015) - (0.015) - (0.015)

Constant 2.141 1.633 2.175 0.893 2.193 1.488 2.174 1.337 2.218 1.662 2.246 2.228 2.311 1.550 2.339 2.608(0.019) (0.424) (0.020) (0.472) (0.019) (0.465) (0.021) (0.530) (0.022) (0.565) (0.022) (0.588) (0.021) (0.612) (0.022) (0.671)

N 1356 1325 1335 1299 1317 1278 1319 1281 1318 1287 1381 1343 1370 1328 1316 1276

1994

1992

2000

NLSY Women born after 1961

Table 3Change in the Black-White Log Wage Gap Induced by Controlling for Schooling-Corrected AFQT for 1990-2000

NLSY Men born after 19611990 1991 1992 1993

1993

1996 1998

g Q g Q g yschooling completed during lifetime. AFQT is a subset of 4 out of 10 ASVAB tests used by the military for enlistment screening and job assignment. It is thesummed score from the word knowledge, paragraph comprehension, mathematics knowledge, and arithmetic reasoning ASVAB tests. All wages are in 1993dollars.

1994 1996 1998 20001990 1991

Figure 13AResidual Black-White Log Wage Gap in 1991 By Grade

At Which We Evaluate The School-Corrected AFQT NLSY79 Males Born After 1961

-0.30

-0.25

-0.20

-0.15

-0.10

-0.05

0.00

8 Or L

ess Year

s

9 Year

s

10 Year

s

11 Year

s

12 Year

s

13-15

Year

Grade At Which We Evaluate the Schooling-Corrected AFQT

Log

Wag

e G

ap

Adjusted Unadjusted

Administrator

Note: We have omitted the results for the "16 or More" category because the low number of minorities in that cell make the correction of the test scores to that schooling level much less reliable than the correction to the other schooling levels. The "unadjusted" line refers to the black-white log wage gap we observe if we do not control for AFQT scores (column I in table 1). Therefore it is a horizontal line since it does not depend on the grade to which we are correcting the test score. The "adjusted" line refers to the black-white log wage gap we observe after we adjust for the AFQT scores corrected to different grades.

Figure 13BResidual Hispanic-White Log Wage Gap in 1991 By Grade

At Which We Evaluate The School-Corrected AFQT NLSY79 Males Born After 1961

-0.12

-0.10

-0.08

-0.06

-0.04

-0.02

0.00

0.02

8 Or L

ess Year

s

9 Year

s

10 Year

s

11 Year

s

12 Year

s

13-15

Year


Log

Wag

e G

ap

Adjusted Unadjusted

Administrator

Note: We have omitted the results for the "16 or More" category because the low number of minorities in that cell make the correction of the test scores to that schooling level much less reliable than the correction to the other schooling levels. The "unadjusted" line refers to the Hispanic-white log wage gap we observe if we do not control for AFQT scores (column I in table 1). Therefore it is a horizontal line since it does not depend on the grade to which we are correcting the test score. The "adjusted" line refers to the Hispanic-white log wage gap we observe after we adjust for the AFQT scores corrected to different grades.

Figure 13CResidual Black-White Log Wage Gap For 1991 By Grade

At Which We Evaluate The School-Corrected AFQT NLSY79 Females Born After 1961

-0.25

-0.20

-0.15

-0.10

-0.05

0.00

0.05

0.10

8 Or L

ess Year

s

9 Year

s

10 Year

s

11 Year

s

12 Year

s

13-15

Year


Log

Wag

e G

ap

Adjusted Unadjusted

Administrator


Figure 13DResidual Hispanic-White Log Wage Gap in 1991 By Grade

At Which We Evaluate The School-Corrected AFQT NLSY79 Females Born After 1961

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

0.18

8 Or L

ess Year

s

9 Year

s

10 Year

s

11 Year

s

12 Year

s

13-15

Year


Log

Wag

e G

ap

Adjusted Unadjusted

Administrator



Black -0.250 -0.144 -0.251 -0.158 -0.302 -0.189 -0.282 -0.182 -0.286 -0.175 -0.373 -0.241 -0.333 -0.175 -0.325 -0.194(0.028) (0.028) (0.028) (0.028) (0.029) (0.028) (0.028) (0.028) (0.031) (0.031) (0.032) (0.031) (0.034) (0.032) (0.035) (0.032)

Hispanic -0.174 -0.056 -0.113 0.005 -0.146 -0.032 -0.159 -0.044 -0.143 -0.018 -0.186 -0.056 -0.195 -0.040 -0.215 -0.070(0.032) (0.032) (0.032) (0.032) (0.033) (0.032) (0.032) (0.032) (0.036) (0.035) (0.038) (0.036) (0.040) (0.037) (0.040) (0.037)

Age - 0.062 - 0.041 - 0.052 - 0.042 - 0.036 - 0.033 - 0.031 - 0.024- (0.014) - (0.014) - (0.014) - (0.014) - (0.016) - (0.016) - (0.017) - (0.016)

AFQT - 0.096 - 0.079 - 0.097 - 0.082 - 0.098 - 0.093 - 0.130 - 0.119- (0.015) - (0.015) - (0.015) - (0.015) - (0.016) - (0.017) - (0.017) - (0.017)

AFQT2 - -0.015 - -0.005 - -0.001 - -0.001 - 0.010 - 0.005 - 0.013 - 0.008- (0.010) - (0.010) - (0.010) - (0.010) - (0.011) - (0.012) - (0.012) - (0.012)

HGC - 0.044 - 0.040 - 0.043 - 0.047 - 0.046 - 0.065 - 0.066 - 0.066- (0.006) - (0.006) - (0.006) - (0.006) - (0.006) - (0.006) - (0.007) - (0.006)

Constant 2.375 0.113 2.372 0.668 2.404 0.281 2.423 0.501 2.458 0.679 2.533 0.540 2.589 0.549 2.629 0.802(0.017) (0.390) (0.017) (0.410) (0.017) (0.422) (0.017) (0.431) (0.018) (0.488) (0.019) (0.531) (0.020) (0.586) (0.020) (0.608)

N 1538 1504 1553 1513 1536 1503 1542 1504 1522 1485 1554 1519 1494 1462 1438 1404


Black -0.172 -0.081 -0.200 -0.101 -0.201 -0.131 -0.167 -0.073 -0.148 -0.069 -0.147 -0.035 -0.201 -0.088 -0.200 -0.086(0.031) (0.030) (0.032) (0.032) (0.031) (0.030) (0.035) (0.033) (0.035) (0.034) (0.035) (0.033) (0.034) (0.032) (0.036) (0.034)

Hispanic -0.003 0.120 -0.017 0.111 -0.059 0.073 0.009 0.139 -0.018 0.118 -0.006 0.137 -0.069 0.091 -0.064 0.096(0.035) (0.033) (0.037) (0.036) (0.036) (0.033) (0.039) (0.037) (0.040) (0.038) (0.041) (0.037) (0.039) (0.036) (0.041) (0.038)

Age - 0.013 - 0.036 - 0.013 - 0.014 - 0.006 - -0.009 - 0.013 - -0.017- (0.015) - (0.016) - (0.015) - (0.017) - (0.017) - (0.017) - (0.017) - (0.017)

AFQT - 0.106 - 0.113 - 0.094 - 0.121 - 0.111 - 0.130 - 0.119 - 0.127- (0.017) - (0.019) - (0.017) - (0.019) - (0.020) - (0.019) - (0.019) - (0.020)

AFQT2 - -0.001 - -0.002 - -0.007 - -0.009 - -0.026 - 0.003 - 0.004 - -0.007- (0.013) - (0.014) - (0.013) - (0.014) - (0.015) - (0.014) - (0.014) - (0.014)

HGC - 0.063 - 0.064 - 0.075 - 0.075 - 0.081 - 0.081 - 0.076 - 0.073- (0.006) - (0.006) - (0.006) - (0.007) - (0.007) - (0.006) - (0.006) - (0.006)

Constant 2.141 0.913 2.175 0.270 2.193 0.784 2.174 0.720 2.218 0.932 2.246 1.397 2.311 0.802 2.339 1.928(0.019) (0.413) (0.020) (0.460) (0.019) (0.443) (0.021) (0.508) (0.022) (0.538) (0.022) (0.558) (0.021) (0.583) (0.022) (0.643)

N 1356 1325 1335 1299 1317 1278 1319 1318 1286 1381 1343 1370 1328 1316 1276

1990 1991 1992 1993 1994

1996

1998

Schooling-corrected AFQT is the standardized residual from the regression of the AFQT score on age at the time of the test dummy variables and final level of schooling completed during lifetime. AFQT is asubset of 4 out of 10 ASVAB tests used by the military for enlistment screening and job assignment. It is the summed score from the word knowledge, paragraph comprehension, mathematics knowledge, andarithmetic reasoning ASVAB tests. All wages are in 1993 dollars. HGC is the highest observed level of schooling completed during the individual's lifetime.

1996

1994

2000

1998 2000

NLSY Women born after 1961

Table 4Change in the Black-White Log Wage Gap Induced by Controlling for Schooling-Corrected AFQT and Highest Grade Completed in 1990-2000

NLSY Men born after 19611990 1991 1992 1993

BLACKS HISPANICS WHITES MALES FEMALES*E1: Enrolled in School 91.75 88.35 93.72 91.67 92.55

(22.79) (26.33) (21.31) (23.32) (22.30)E1: Working for pay > 20hs wk and enrolled 62.77 60.00 59.11 60.84 58.86

(32.44) (31.02) (33.43) (32.00) (33.65)E1: Working for pay > 20hs wk and not enrolled 77.01 76.39 83.11 80.09 79.78

(31.44) (30.18) (27.58) (28.88) (29.95)E1: Got pregnant 7.99 7.78 4.68 0.00 6.32

(20.43) (17.64) (12.99) 0.00 (16.53)E1: Got someone pregnant 13.52 12.33 6.27 9.46 0.00

(21.75) (20.65) (14.77) (18.39) 0.00E1: Seriously drunk at least once 11.64 20.37 24.12 21.55 18.00

(23.77) (29.78) (33.81) (32.06) (29.44)E1: Victim of violent crime 16.16 16.11 13.36 15.40 13.97

(22.99) (21.92) (18.89) (21.44) (20.04)E1: Arrested 12.14 11.43 8.74 13.73 6.65

(20.55) (19.49) (16.08) (20.69) (14.33)E1: Dead from any cause 22.53 19.02 16.56 17.55 19.75

(25.35) (23.00) (20.36) (21.92) (23.03)**E2: Received a high school diploma 92.99 88.92 95.44 92.33 94.57

(19.45) (23.64) (15.57) (19.63) (17.58)E2: Served time in jail or prison 5.03 7.53 4.54 7.18 3.40

(13.00) (16.35) (11.76) (15.20) (10.41)E2: Mothered or Fathered a baby 21.20 21.00 14.86 19.22 16.28

(29.34) (27.69) (23.14) (25.63) (26.30)E2: Dead from any cause 22.27 21.48 19.01 19.60 21.08

(24.73) (23.60) (20.72) (22.20) (22.75)***E3: Earned a 4-yr college degree 74.22 66.93 73.75 68.75 76.90

(31.44) (31.59) (31.50) (32.02) (30.37)E3: Working for pay > 20hs wk 97 90.48 89.72 94.39 92.81 91.85

(20.05) (19.23) (13.71) (16.05) (17.93)P1: Enrolled in School 90.68 89.39 94.03 90.80 93.64

(23.70) (25.08) (20.07) (23.80) (20.48)P1: Working for pay > 20hs wk and enrolled 51.42 50.52 42.65 47.14 45.09

(34.84) (37.10) (38.16) (37.05) (37.54)P2: Received a high school diploma 91.51 91.94 96.19 93.12 95.13

(21.14) (19.97) (15.10) (19.43) (16.31)P2: Served time in jail or prison 4.77 3.87 2.62 4.84 2.06

(14.35) (12.99) (9.39) (13.93) (8.76)P2: Mothered or Fathered a baby 17.55 19.32 12.58 15.54 14.94

(28.84) (27.81) (21.55) (24.66) (25.68)P3: Earned a 4-yr college degree 68.29 67.15 69.85 65.68 72.77

(33.44) (32.35) (32.29) (34.03) (30.63)P3: Working for pay > 20hs wk 97 93.22 92.89 94.95 95.87 92.18

(16.17) (17.82) (13.30) (12.93) (16.96)

TABLE 5 - MEANS AND ST DEV OF EXPECTATIONS

NLSY97 ROUND1BY RACE AND SEX

* E1 means expectations a year from interview date. P1 means expectations of youth's parent a year from interview date.** E2 means expectations by youth's 20th birthday. P2 means expectations of youth's parent by youth's 20th birthday.*** E3 means expectations by youth's 30th birthday. P3 means expectations from youth's parent by youth's 30th birthday.1In round 1, respondents who were born in 1980 or 1981 were surveyed on their beliefs about the future. Asked to assess the probability that certain events would occur in a specified time period, the respondents were instructed to use a scale from 0 (impossible) to 100

BLACKS ACTUAL HISPANICS ACTUAL WHITES ACTUALAll individuals 0.912 0.734 0.881 0.717 0.934 0.790

(0.232) (0.442) (0.265) (0.451) (0.219) (0.407)Individuals Enrolled in 1997 0.936 0.764 0.915 0.758 0.957 0.819

(0.188) (0.425) (0.217) (0.429) (0.173) (0.385)

Table 6ATable of Means for Males By Race

Expectation1 Of Being Enrolled Next YearNLSY97 Round 1

1In round 1, respondents who were born in 1980 or 1981 were surveyed on their beliefs about the future. Asked to assess the probability thatcertain events would occur in a specified time period, the respondents were instructed to use a scale from 0 (impossible) to 100 (certain).

BLACKS ACTUAL HISPANICS ACTUAL WHITES ACTUALAll individuals 0.885 0.734 0.880 0.717 0.930 0.790

(0.255) (0.442) (0.259) (0.451) (0.217) (0.407)Individuals Enrolled in 1997 0.909 0.764 0.911 0.758 0.954 0.819

(0.221) (0.425) (0.220) (0.429) (0.169) (0.385)

Table 6BTable of Means For Males by Race

Parental Expectation1 of Youth Being Enrolled Next YearNLSY97 Round1

1In round 1, respondents who were born in 1980 or 1981 were surveyed on their beliefs about the future. Asked to assess the probability thatcertain events would occur in a specified time period, the respondents were instructed to use a scale from 0 (impossible) to 100 (certain).

0

.2

.4

.6

Fra

ctio

n

Male FemaleHispanic Black White Hispanic Black White

The height of the bar is produced by dividing the number of people who report falling in a particulareducational expectation cell by the total number of people in their race-sex group.

Children of NLSY79

Figure 14AChild’s Own Expected Educational Level at Age 10

by Race and Sex

Drop Out High School Graduate

Some College or 4−Year College Graduate More than 4−Year College

0

.2

.4

.6

.8

Fra

ctio

n


The height of the bar is produced by dividing the number of people who report falling in a particulareducational expectation cell by the total number of people in their race−sex group.

Children of NLSY79

Figure 14BMother’s Expected Educational Level For the Child at Age 6

by Race and Sex



0

.1

.2

.3

.4

.5

Fra

ctio

n


The height of the bar is produced by dividing the number of people who report falling in a particulareducational expectaion cell by the total number of people in their race−sex group.

Children of NLSY79

Figure 14CYoung Adult’s Own Expected Educational Level at Age 17

by Race and Sex



35

40

45

50

Ave

rage

Per

cent

ile S

core

4−5 6−7 8−9 10−11 12−13 14Child’s Age


Mothers were asked 28 age−specific questions about frequency, range and type of specific behavior problems thatchildren age four and over may have exhibited in the previous three months. Factor analysis was used to determinesix clusters of questions. This test is one such cluster. The responses for each cluster were dichotomized and summedto produce a raw score. The percentile score was then calculated separately for each sex at each age from the raw scoreA higher percentile score indicates a higher incidence of problems.


Figure 15APercentile Antisocial Behavior Score By Race and Age Group

30

35

40

45

50

Ave

rage

Per

cent

ile S

core

4−5 6−7 8−9 10−11 12−13 14Child’s Age


Mothers were asked 28 age−specific questions about frequency, range and type of specific behavior problems thatchildren age four and over may have exhibited in the previous three months. Factor analysis was used to determinesix clusters of questions. This test is one such cluster. The responses for each cluster were dichotomized and summedto produce a raw score. The percentile score was then calculated separately for each sex at each age from the raw scoreA higher percentile score indicates a higher incidence of problems.


Figure 15BPercentile Antisocial Behavior Score By Race and Age Group

38

40

42

44

46

Ave

rage

Per

cent

ile S

core

4−5 6−7 8−9 10−11 12−13 14Child’s Age


Adjusted by permanent family income, mother’s education and age−corrected AFQT, and home score.Adjusted indicates that we equalized the family background characteristics across all race groups by setting them atthe mean to purge the effect of family environment disparities. Permanent income is constructed by taking the averageof annual family income discounted to child’s age 0 using a 10% discount rate. Age−corrected AFQT is the standardizedresidual from the regression of the raw AFQT score on age at the time of the test dummy variables. Home score is anindex of quality of the child’s home environment.


Figure 16AAdjusted Percentile Antisocial Behavior Score By Race and Age Group

30

35

40

45

Ave

rage

Per

cent

ile S

core

4−5 6−7 8−9 10−11 12−13 14Child’s Age


Adjusted by permanent family income, mother’s education and age−corrected AFQT, and home score.Adjusted indicates that we equalized the family background characteristics across all race groups by setting them atthe mean to purge the effect of family environment disparities. Permanent income is constructed by taking the averageof annual family income discounted to child’s age 0 using a 10% discount rate. Age−corrected AFQT is the standardizedresidual from the regression of the raw AFQT score on age at the time of the test dummy variables. Home score is anindex of quality of the child’s home environment.


Figure 16BAdjusted Percentile Antisocial Behavior Score By Race and Age Group

nber working paper series labor market …

Documents